3?
CONTROL OF SULFUR EMISSIONS
FROM OIL SHALE RETORTING USING
SPENT SHALE ABSORPTION
PILOT PLANT TESTING
Kenneth D. VanZanten
Gerald R. Chiaramonte
J 4 A Associates, Inc.
18200 West Highway 72
Golden, Colorado 80401
Edward R. Bates
U.S. Environmental Protection Agency ;
A1r and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
NOTICE
The research described 1n this article has been funded wholly
or 1n part by the United States Environmental Protection Agency
through contract 68-03-1969 to 0 Ğ A Associates, Inc. It has
been subjected to Agency review and 1s approved for publication.
A. BACKGROUND
Control of sulfur emissions constitutes a major
.portion of the environmental control cost for oil
shale facilities. For example, the Denver Research
Institute estimated costs (in 1980 dollars) In the
range of $1 to $3 per barrel of shale oil
produced'1,2,3,). These substantial sulfur control
costs have encouraged developers to seek less costly
9-
but equally or more effective methods for limiting
sulfur emissions. Recently, a strong industry trend
has been to look toward the potential for combusting
carbonaceous retorted shale to recover Its energy
value (a plus In terms of economics and resource
conservation), while exploring the possibility of
ibsorblng the sulfur gases produced during retorting"
onto the calcined carbonate material present after
combustion of retorted western oil shale.
EPA awarded a contract to J S A Associates to
investigate the environmental advantages/
disadvantages of absorbing S02 onto combusted
retorted oil shale. The objective of this program
was to obtain more information in support of its PSD
(Prevention of Significant Deterioration) permitting
decisions on sulfur control and to investigate
whether emission of other pollutants such as nitrogen
oxides (NOX) and trace elements might be ;
significantly Increased 1n the process. The program
was done 1n two phases. Phase I developed an
engineering assessment and costs for application of
this sulfur absorption process to selected Reading
? .'
retorting processes. In Phase II, experimental work
1n an .integrated oil shale pilot plant defined
operabllity and proof of principle and defined trace
i
element emissions. ;
I
B. THE ASSP CONCEPT I
;
The ability of combusted carbonate-containing spent
shale to absorb S02 gives rise to a novel concept for
controlling sulfur emissions 1n oil shale plants.
This concept will be referred to as ASSP which stands
for Absorption on Spent Shale Process. '
-------
The ASSP concept has several potential advantages
over conventional sulfur removal technologies:
o The sorbent Is cheap and Inherently
abundant 1n oil shale plants.
o The process requires combustion of the
spent shale which is already Incorporated
Into several of the retorting technologies
or which would be a useful add-on to
recover residual carbon values.
o SI nee non-HgS compounds are converted to
S02 by combustion. ASSP could represent a
more efficient removal relative to gas
sweetening processes which only remove HgS.
+
The ASSP concept uses a fluldlzed transport system to
combust either raw or retorted shale, thereby
providing the vehicle for converting sulfur compounds
to S02 and absorbing the SO? In the shale matrix.
The concept envisions either a conventional
dense-phase fluldlzed bed or a dilute-phase contactor
(lift pipe). Key elements of the process are shown
In Figure 1.
C. PHASE I; CONCEPTUAL DESIGN AND ECONOMICS
For evaluation purposes, specific projects were
chosen as representative of the three retort types:
o Direct heated - Modified In-:S1tu (HIS) witt
Unishale C - Cathedral Bluffs
o Indirect heated - Unishale B( - Union 011
o Integral Combustor - Lurgl - R1o Blanco
- Un1shal,e C - Union Oil
This study assumed that MDEA (MethyldletJianolamine)
absorption 1s used to remove acid gase,s from Indirect
heated retort gases and that regenerated add gases
are burned 1n the ASSP cpmbustor. HIS gases were
assumed to be processed 1n the ASSP combustor without
pretreatment.
For comparison purposes, conventional sulfur removal
processes were evaluated:
o Direct heated - Case A:' Unlsulf +
Flue Gas
Desulfurizatlon on HIS
gases
Case B:i Unlsulf +
Stretford on MIS gases
o Indirect heated - Unlsulf!
o Integral Combustor - DEA + Stretford on
Lurgi !
- Unlsulf|on Unishale C
TABLE 1. COST COMPARISON FOR ASSP
Retort Type
Retorting Process
ASSP Incremental
Capital Cost, $106
ASSP Incremental Annual
Operating Cost, $106/yr.
Plant Capacity TPSD
(kg/sec)
Direct Heated Indirect
Case A, Case B Heated
HI S/Un1 shale C Unishale B
-71.2 63.2 +90.2
+10.83 +12.07 -19.21
36,200 13,600 27,200
(380) (143) (286)
Integral
Conbustor
Lurgl Uni shale C
-13.0 -32.1
-2.29 -1.56
119,000 27,200
(1251)
(286)
Source: Reference (4).
TPSD: Tons Per Strea'm Day.
-------
o
o
ce
ce
ZD
H-
oo
I
I
I
L
§1
o. I
o:
HJQ cc
-JUJUJ
iui-
ou- <
CO Z
a
UJ
a:
- ui (
ou.
-------
Major equipment costs were taken from EPA Pollution
Control Technical Manuals (PCTMsjd.Z). ASSP
equipment was sized and costs factored from in-house
data and PCTMs. Costs were factored to first quarter
1985.
A discussion of the Phase I work 1s given 1n more
detail in a previous paper.(^)
The test facility used 1n Phase II was a'pilot plant
built by Tosco Corporation to develop their
Hydrocarbon Solids Process (HSP)'5'. The pilot plant
has a nominal capacity of 6 tons per day'(63 g/sec)
of oil shale and contains a fluldized bed combustor
which 1s 18 1n. (0.46m) in diameter. Figure 2 is a
|
process flow diagram of the plant. A description of
the process 1s given below.
Results of the cost study .showed changes 1n
incremental capital and operating costs for ASSP
relative to conventional processing in Table 1. .
These cost comparisons show that the best potential
for application of ASSP are .those processes which
already have a spent shale combustor Integrated Into
the retorting process (e.g., Lurgl, Unishale C,
Chevron STB, and Tosco HSP). Capital and operating
cost savings for Unishale C and Lurgi are primarily a
result of deleting the Unlsulf and Stretford plants.
r
Economics for the Indirect and direct heated retorts
are good to marginal. Factors which will affect the
economics are:
o How effectively combustor heat can be
utilized (simple steam raising is the least
desirable).
o The value of steam.
o The use of fast or circulating fluid beds
to reduce investment in combustor equipment.
D. PHASE II: PILOT PLANT TESTING
Raw oil shale, crushed to minus 1/4-in. (0.0064m) and
smaller, Is pneumatically lifted to the shale feed
weigh hopper system from which shale Is metered Into
the retort at a constant rate. The raw s'hale from
the weigh hopper is preheated up to 300 to 500'F
(421-533K). The retort 1s an Inclined rotating
cylinder in which oil shale and hot heaticarrier
solids (from the fluid bed combustor) are mixed. The
mixture of heat carrier and oil shale is conveyed
concurrently through the retort to the retort
accumulator. The feed rates of raw oil shale and
i
heat carrier are adjusted to maintain the desired
temperature 1n the retort, approximately J900"F (755K).
i
The mixture of spent shale and heat carrier from the
retort, called retorted solids, is pneumatically
conveyed from the accumulator discharge screw into
the fluid bed combustor using superheated steam. The
fuel residue on the spent shale (primarily organic
carbon and hydrogen) Is combusted to provide part or
all of the heat required to pyrolyze the oil shale.
Combusted solids', which consititute the heat carrier,
are drawn off from the fluid bed combustor and are
recycled to the retort. i
1. Description of the Pilot Plant
-------
İ
j 1
r* K ?
- \ r< .
- = /
3S /
s *
" 5*
ii- <_> n
u
Ğ
a
4J
J
£
i
|
^
01
M
HI
O
O
H
O
i-l
H
(VI
0)
3,
-------
The combustor 1s an atmospheric, dense-phase,
bubbling, fluldized bed. The spent shale fuel 1s
supplemented as needed by Injection of natural gas or
retort gas Into the bed. Solid fuels such as raw oil
shale can also be used. The bed 1s fluldlzed by air
and/or hot flue gases from an external burner. Flow
rates of air and flue gas are adjusted to maintain
bed fluldlzatlon, bed temperature, and oxygen
concentration 1n the combustor flue gas.
Flue gas and entrained shale ash from the combustor
are cooled 1n a heat exchanger, and the ash 1"s
separated from the flue gas 1n a baghouse. From the
baghouse weigh bin the ash flows to a moisturizer
t
where 1t 1s mixed with water prior to disposal. The
clean flue gas flow Is measured with an orifice meter
prior to being vented to the atmosphere.
o How effective Is ASSP In;controlling su
emissions? j
o Will ASSP produce large quantities of N
o What are the most favorable operating
conditions to achieve maidnum sulfur
control while holding NOX emissions to
minimum?
o Will retorted or raw oil shale combusti
produce significant emissions of trace
elements such as mercury or cadmium?
3. Experimental Procedure '
Parachute Creek oil shale obtained jfrom the Colon;
mine was used 1n the pilot plant program. This si
was crushed to minus l/4-1n. (0.0064m) particle si
and had a nominal richness of 34-37 gal. per ton
(142-154 L/Tonne). The shale used;1s similar to
shale being processed by the Union iflil coumercial
plant. This shale has significant amounts of calc
and magnesium carbonates, which (whjen decomposed 1
the oxides) are available for sulfur absorption.
Pyrolysis vapors from the retort are cooled and the
oil and water condensed 1n a quench tower and
overhead condenser. The non-condensed retort vapors
are either metered and sent to a flare or are
diverted to the fluid bed combustor through a blower
used to overcome the pressure In the bed. For the
majority of the pilot plant tests, retort gas was
burned in the fluid bed to supply the H2S and non-HzS
sulfur compounds. In addition, H2S and COS from
pressurized cylinders were used to "spike" the retort
gas to allow significantly higher sulfur
concentrations 1n the Injected retort gas than would
have been possible with only retort gas.
2. Test Objectives
Key questions addressed In the Phase II test program
included:
The pilot plant was operated for 10 days between
October 14 and 24, 1985. A total of 44 "tests" wĞ
conducted during which plant operating data were
recorded. j
Some of the key process variables evaluated in the
pilot plant program were: ;
o Bed temperature \
o Solids residence time (bed depth and sol
circulation rate)
o Gas residence time (superficial velocity
o Ca/S mole ratio
o Flue gas oxygen concentration
o Raw shale/spent shale ratio
o Single stage and two stage combustion
During single stage combustion tests, all combust1<
air flowed through the fluid bed and superheated
steam was used to pneumatically convey retorted
solids to the combustor via the transfer line. In
this mode, the bed was normally oxygen-rich.
-------
During two stage combustion tests, combustion air to
the bed was reduced until the flue gas oxygen
concentration fell to zero. Then, overflre air was
added to the retorted solids transfer line while
simultaneously reducing superheated steam flow until
the desired flue gas 02 level was achieved. Since
the transfer line does not enter the fluid bed,
overflre air and superheated steam do not pass
through the bed. Thus overflre air can combust CO
and trace hydrocarbons In only the freeboard portion
of the combustor.
4. Discussion of Results
The range of key operating conditions for the 44
tests performed are summarized In Table 2.
Over this range of conditions, flueigas composltloi
and organic carbon combustion efficiency ranged as
shown In Table 3. :
To Increase the sulfur concentration 1n the retort
i
gas, HjS and COS from pressurized cylinders were
Injected Into the retort gas upstream of the sampl-
point. This Increased the HgS concentration from e
Initial concentration of about 0.4 - 0.5 vol % to 2
4 vol * and 1n some tests, nearly 10 vol I. The
retort gas was spiked with HgS and COS 1n 33 of the
44 tests.
5. Process Variable Correlations
Correlations of emissions (SOg, NOX; CO, trace HC)
TABLE 2. RANGE OF OPERATING CONDITIONS AND PROCESS VARIABLES
Bed Temperature, *F (K)
Freeboard Temperature, *F (K)
Retorted Solids to Combustor, Ib/hr (g/sec)
Raw Shale to Combustor, Ib/hr (g/sec)
Retort Gas to Combustor, scfm (Nm^/sec)
HjS 1n Retort Gas, vol $
Bed Depth, ft (m)
Sol Ids Residence Time, m1n.
Gas Superficial Velocity, ft/sec (m/sec)
Gas Residence Time, sec
Flue Gas Oxygen, vol I
Carbonate Decomposition, %
Ca/S Mole Ratio
1127-1558
1273-1593
2487-3615
0- 133
0-6.66
0.43-9.28
3.27-4.28
8.07-18.72
3.78-7.20
0.46-1.13
0-6.25
45.5-83.3
6.20-10.25
(881-1121)
(962-1140)
(315-458)
( 0-17)
( 0-11.3)
( 1-1.3)
(1.15-2.19)
lABLt 3. KANlit
-------
with key process variables indicated that the only
significant factor which affected emissions was flue
gas oxygen concentration. Smoothed curves of the
experimental data are shown in Figures 3 and 4. Note
that the NOX curve is presented as a band reflecting
substantial data scatter.
Key findings of the test program were:
o S02 emissions were easily controlled to low
levels at virtually all conditions tested,
probably as a result of the high Ca/S
ratios used.
Thus the Inlet sulfur corcentraion 1s
Immaterial providing the Ca/S ratio Is
adequate.
o Reasonably good NOX control could be
obtained with flue gas oxygen
concentrations below about 3 vol 1. The
lowest NOX concentrations were seen at Oj
levels approaching zero but at the expense
of higher CO and trace hydrocarbon
emissions.
o Good control of CO and trace hydrocarbon
, emissions could be obtained at 03 levels
above about 2 vol $.
Emissions of NOX move in a direction opposite to SOj,
CO, and trace hydrocarbon emissions. Thus, finding a
set of operating conditions which minimize all four
represents a compromise. One test was run which
produced nearly optimum results. Conditions for this
test were:
Bed Temperature 1227*R (937K)
Solids Residence Time 9.4 roin.
Gas Residence Time 0.9 sec
Gas Superficial Velocity 4.4 ft/sec (1.3 m/se<
Flue Gas Og 2.6 volt
Ca/S Mole Ratio 10.3 '
Raw Shale/Spent Shale Ratio , 1:36
At these conditions the following results were
obtained: !
S02 11 ppmv
MOX 160 ppmv
CO 0.27 yoU
Trace Hydrocarbon 388 ppmv
Combustion Efficiency 89% !
6. Design Recommendations !
Based on the pilot plant data obtained in this study.
fluid bed operating conditions are reconmended to
optimize SO? and NOX control. In general, condition:
that favor low SOg emissions also favor low CO and
trace hydrocarbon emissions but do not favor low NOX
emissions. The general ranges of operating
conditions which produced reasonable results from
both an operating and emissions viewpoint are given
below. Conditions used 1n the Phase ! conceptual
design work are shown for comparison 1n Table 4.
This comparison Indicates that the conditions chosen
for the conceptual design are reasonable and in most
i
cases conservative. |
TABLE 4. RECOMMENDED FLUID BED COMBUSTOR OPERATING CONDITIONS
Operating Conditions '
Fluid Bed Temperature, *F (K)
Solids Residence Time, rain
Gas Residence Time, sec
Gas Superficial Velocity, ft/sec
(m/sec)
Flue Gas Oxygen, vol *
Carbonate Decomposition, %
Ca/S Hole Ratio
Raw Shale/Spent Shale Ratio
Reconmended
1150-156U (B3/-lllb)
11-14
0.5-1.0
7+ (2.1+)
2+
45+
6+
3/97
Conceptual Design
1350 (1UU5)
14
1.0 :
5.0 ( 1.5)
3.0 ',
60
23
7/93 '
-------
2
0.
Q_
4O
35 -
30 -
25 -
2O -
15 -
10 -
5 -
O 2 4
FLUE GAS O2.
FIGURE 3. EFFECT OF FLUE GAS OXYGEN ON S02 AND NOX EMISSIONS.
7OO
- eoo
- 5OO
- 400
: Q.
0-
X
: i
- 300
- 2OO i
- 1OO
ui
3
O '
o
O
LU
FLUE GAS 02, VOL%
FIGURE ii. EFFECT OF FLUE GAS OXYGEN ON CO AND TRACE HYDROCARBON EMISSIONS.
-------
7. Trace Element Sampling and Analysis
Selected process streams were sampled and analyzed
for the following trace elements: arsenic, cadmium,
mercury, beryllium, lead and fluorine (as fluoride).
The streams sampled and analyzed were raw shale,
retorted solIds, heat carrier, baghouse ash, retort
gas, and combustor flue gas. Oil and retort water
were not analyzed. The primary goal was to determine
the trace element concentration In the retort gas and
flue gas.
Previous 1nvest1gators<6-l2) measured trace element
concentrations 1n various process streams from
laboratory and simulated In-sltu retorts. Mercury
concentrations of from less than 0.2 to 8,200 ug/m3
have been reported in retort gases. Cadmium
concentrations of from 1 to over 1000 ug/m3 have been
reported in retort gases. Arsenic concentrations
from 5 to 155 ug/m3 have been reported in gas streams
from oil shale retorts. No data have been reported
on lead, beryllium, and fluoride In oil shale
processing gas streams.
Retort gas and combustor flue gas were sampled during
three pilot plant tests: Tests 7, 12, and 19C.
Retorting and combustion temperatures were varied for
these three tests; average temperatures' are given:
Test
Number
Average
Retorting
Temp, *F (K)
1015 (FT?)
930 (772)
860 (733)
Average
Combustion Temp,
*F (K)
1550 (1116)
12 930 (772) 1430 i (1050)
19C 860 (733) 1240 ', ( 944)
During Test 12, an aqueous spike solution containing
3 g/L Hg and 3 g/L Cd (prepared from the nitrate
r
salts) was pumped into the bottom of the combustor to
determine their fate 1n the combustor. [The amount of
mercury and cadmium fed In the spike represents about
4700 times and 170 times, respectively,'of the
amounts of mercury and cadmium entering,the system In
i
the raw shale during the 2-hour spike period.
A summary of analytical results is given 1n Table 5.
Table 6 gives the percentage of trace elements
present in the raw shale feed which was 'found In the
retort gas and flue gas. In Test 12 the! amount of
mercury and cadmium added to the combustor is
Included as part of the total. Note that, although
the trace elements were fed to the combustor,
significant amounts of mercury (106 m1crograms/m3)
were found 1n the retort gas. Mercury and possibly
some cadmium were probably deposited on the heat
carrier in the combustor and recycled to the retort
where they were re-volatilized.
TABLE 5. TRACE ELEMENT ANALYTICAL RESULTS
Test
7
12
19C
Stream
Retort Gas
Flue Gas
Retort Gas
Flue Gas
Retort Gas
Flue Gas
Concentration, micrograms/m3
Hg
4
4
106
24,720
4
22
Co
184
26
264
25
2
15
AS
115,355
8
88,103
9
49
9
PD
2,546
113
146
6
9
9
Be
2
9
2
2
2
2
F
35
35
35
35
35
35
10
-------
TABLE 6. PERCENTAGE OF TRACE ELEMENTS PRESENT IN RAW SHALE
FOUND IN RETORT GAS AND FLUE GAS
Test 7
Hg
Cd
As
Pb
Be
F
Test 12
Hg(a)
Cd(a)
As
Pb
Be
F
Test 19C
Hg
Cd
AS
Pb
Be
F
I Found In
Retort Gas
NO
2
13.8
0.65
ND
NO
0.01
0.02
11.2
0.03
ND
ND
* Found In
Retort Gas
(b)
(b)
0.004
0.001
ND
NO
% Found In
Flue Gas
ND
7 !
0.03
0.8
2.7
ND :
41.3
0.005
0.03 :
0.04 :
ND
ND :
% Found In
Flue Gas ',
(b) ;
(b) :
0.03 ;
0.7
NO
ND :
ND - Not detected
(a) ğ Includes metal spikes
(b) - Not applicable due to unknown amount of Hg and Cd still present from
the spiking of these metals 1n Test 12
Results of the trace element tests Indicated some
relative trends with regard to emissions but because
of the short duration of the sampling, no hard
conclusions can be reached which would allow
extrapolation of results to long term steady-state
operations. Some of the key observations were:
o Lead, beryllium and fluoride were found to
have low volatility. That 1s, of the
amounts present 1n raw shale, only very
small percentages were volatilized to the
gas streams.
Arsenic was found 1n significant
concentrations 1n the retort gas (100-400
i
ppmv), although the amount of arsenic
represented less than 15% of that 1n the
raw shale. ,
So little mercury was present 1n the raw
shale that mercury emissions could not be
characterized with high accuracy. Mercury
emissions were very low except during the
spike, Indicating that mercury, If present
In higher concentrations In ;the raw shale,
could possibly pose emissions problems.
11
-------
o Cadmium demonstrated moderate volatility at
higher retort and combustor temperatures
but emissions represented less than 10% of
cadmium present in raw shale.
There is some evidence that mercury and cadmium
introduced to the combustor during the spike test
condensed within the retort equipment and
revolatflized over time. . However, because of the
limited number of samples taken, It would not be
prudent to draw any hard conclusions. Longer term
steady-state operations would have to be studied to
determine the fate of mercury and cadmium with,more
certainty.
REFERENCES
1. Denver Research Institute, Pollution Control
Technical Manual for Lurgl 011 Shale Retorting
with Open Pit Mining, EPA-600/8-83-005, NTIS
PB83-200204, April 1983.
2. Denver Research Institute, Pollution Control
Technical Manual for Modified In-SItu Oil Shale
Retorting Combined with Lurgl Surface Retorting,
EPA-600/8-83-004. NTIS PB83-200121, April 1983.
3. Denver Research Institute, Pollution Control
Technical Manual for Tosco II 011 Shale
Retorting with Underground Mining,
EPA-600/8-83-003, NTIS PB 83-200212, April 1983.
4. VanZanten, K.D., et al, "Control of Sulfur
Emission From 011 Shale Retorting Using Spent
Shale Absorption," AIChE 1985 Annual Meeting,
Chicago, IL, November 1985.
5. Hall, R.N., "Hydrocarbon Solids Process-HSP
Technology," AIChE Spring National Meeting,
Anaheim, CA, June 1982. t
6. Fox, J.P., "Distribution of Mercury During
Simulated In-Situ 011 Shale Retorting,"
Environmental Science and Technology, 1_9, pp
316-322, April 1985. '
i
7. Hodgson, A.T., et al. "Mercury Mass Distribution
During Laboratory and Simulated In-Situ Oil
Shale Retorting," Lawrence Berkeley Laboratory,
Berkeley. CA, February 1982, Report LBL-12908,
I
8. Hodgson, A.T., Pollar, M.J., Brown, N.J.,
"Mercury Emissions From a Modified In-Situ Oil
Shale Retort," Atmos. Environ. 1_8, pp. 247-253,
1984. |
9. 01 sen, K.B., "Characterization of Mercury,
Arsenic and Selenium in the Product Streams of
the Pacific Northwest Laboratory 6rKg, Retort,"
Richland, WA, Prepared for the U.S: DOE,
Contract DE-AC06-76RL01830, July 1985.
10. Fruchter, J.S., et al, "The Partitioning in
Aboveground Oil Shale Retort Pilot Plant."
Environmental Science and Technology, November
1980. I
11. Fox, J.P., et al, "The Partitioning of As, Cd,
Cu, Hg, Pb and Zn During Simulated iln-Situ Oil
Shale Retorting," 10th Annual Oil Shale
Symposium, Golden, CO, 1977. !
12
-------
12. Fox, J.P, et al, "Partitioning of Major, Minor
and Trace Elements During Simulated In-S1tu 011
Shale Retorting in a Controlled State Retort,"
12th Annual Oil Shale Symposium, Golden, CO,
1979.
13
------- |